MASDF-Net: A Multi-Attention Codec Network with Selective and Dynamic Fusion for Skin Lesion Segmentation DOI Creative Commons

Jinghao Fu,

Hongmin Deng

Sensors, Journal Year: 2024, Volume and Issue: 24(16), P. 5372 - 5372

Published: Aug. 20, 2024

Automated segmentation algorithms for dermoscopic images serve as effective tools that assist dermatologists in clinical diagnosis. While existing deep learning-based skin lesion have achieved certain success, challenges remain accurately delineating the boundaries of regions with irregular shapes, blurry edges, and occlusions by artifacts. To address these issues, a multi-attention codec network selective dynamic fusion (MASDF-Net) is proposed this study. In network, we use pyramid vision transformer encoder to model long-range dependencies between features, innovatively designed three modules further enhance performance network. Specifically, (MAF) module allows attention be focused on high-level features from various perspectives, thereby capturing more global contextual information. The information gathering (SIG) improves skip-connection structure eliminating redundant low-level features. multi-scale cascade (MSCF) dynamically fuses different levels decoder part, refining boundaries. We conducted comprehensive experiments ISIC 2016, 2017, 2018, PH2 datasets. experimental results demonstrate superiority our approach over state-of-the-art methods.

Language: Английский

Fast co-clustering via anchor-guided label spreading DOI
Fangyuan Xie, Feiping Nie, Weizhong Yu

et al.

Neural Networks, Journal Year: 2025, Volume and Issue: 185, P. 107187 - 107187

Published: Jan. 22, 2025

Language: Английский

Citations

0
U3UNet: An accurate and reliable segmentation model for forest fire monitoring based on UAV vision DOI
Hailin Feng, Jiefan Qiu, Long Wen

et al.

Neural Networks, Journal Year: 2025, Volume and Issue: 185, P. 107207 - 107207

Published: Jan. 30, 2025

Language: Английский

Citations

0
Uncertainty-Aware Adaptive Multiscale U-Net for Low-Contrast Cardiac Image Segmentation DOI Creative Commons
A. S. M. Sharifuzzaman Sagar, Muhammad Zubair Islam, Jawad Tanveer

et al.

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(4), P. 2222 - 2222

Published: Feb. 19, 2025

Medical image analysis is critical for diagnosing and planning treatments, particularly in addressing heart disease, a leading cause of mortality worldwide. Precise segmentation the left atrium, key structure cardiac imaging, essential detecting conditions such as atrial fibrillation, failure, stroke. However, its complex anatomy, subtle boundaries, inter-patient variations make accurate challenging traditional methods. Recent advancements deep learning, especially semantic segmentation, have shown promise these limitations by enabling detailed, pixel-wise classification. This study proposes novel framework Adaptive Multiscale U-Net (AMU-Net) combining Convolutional Neural Networks (CNNs) transformer-based encoder–decoder architectures. The introduces Contextual Dynamic Encoder (CDE) extracting multi-scale features capturing long-range dependencies. An Feature Decoder Block (AFDB), leveraging an Attention (AFAB) improves boundary delineation. Additionally, Spectral Synthesis Fusion Head (SFFH) synthesizes spectral spatial features, enhancing performance low-contrast regions. To ensure robustness, data augmentation techniques rotation, scaling, flipping are applied. Laplacian approximation employed uncertainty estimation, interpretability identifying regions low confidence. Our proposed model achieves Dice score 93.35, Precision 94.12, Recall 92.78, outperforming existing

Language: Английский

Citations

0
CTDUNet: A Multimodal CNN–Transformer Dual U-Shaped Network with Coordinate Space Attention for Camellia oleifera Pests and Diseases Segmentation in Complex Environments DOI Creative Commons

Ruitian Guo,

R. Y. Zhang, Hao Zhou

et al.

Plants, Journal Year: 2024, Volume and Issue: 13(16), P. 2274 - 2274

Published: Aug. 15, 2024

is a crop of high economic value, yet it particularly susceptible to various diseases and pests that significantly reduce its yield quality. Consequently, the precise segmentation classification diseased Camellia leaves are vital for managing effectively. Deep learning exhibits significant advantages in plant pests, complex image processing automated feature extraction. However, when employing single-modal models segment

Language: Английский

Citations

1
MASDF-Net: A Multi-Attention Codec Network with Selective and Dynamic Fusion for Skin Lesion Segmentation DOI Creative Commons

Jinghao Fu,

Hongmin Deng

Sensors, Journal Year: 2024, Volume and Issue: 24(16), P. 5372 - 5372

Published: Aug. 20, 2024

Automated segmentation algorithms for dermoscopic images serve as effective tools that assist dermatologists in clinical diagnosis. While existing deep learning-based skin lesion have achieved certain success, challenges remain accurately delineating the boundaries of regions with irregular shapes, blurry edges, and occlusions by artifacts. To address these issues, a multi-attention codec network selective dynamic fusion (MASDF-Net) is proposed this study. In network, we use pyramid vision transformer encoder to model long-range dependencies between features, innovatively designed three modules further enhance performance network. Specifically, (MAF) module allows attention be focused on high-level features from various perspectives, thereby capturing more global contextual information. The information gathering (SIG) improves skip-connection structure eliminating redundant low-level features. multi-scale cascade (MSCF) dynamically fuses different levels decoder part, refining boundaries. We conducted comprehensive experiments ISIC 2016, 2017, 2018, PH2 datasets. experimental results demonstrate superiority our approach over state-of-the-art methods.

Language: Английский

Citations

1